Q-learning-based multi-agent initiative recommendation method for agriculture capital electronic commerce

Abstract

The invention discloses a Q-learning-based multi-agent initiative recommendation method for agriculture capital electronic commerce. The Q-learning-based multi-agent initiative recommendation method is characterized by including the steps of 1), multi-agent initiative recommendation system structuring; 2), state definition; 3), initialization; 4), Q-learning calculation and recommendation direction selection; 5), recommendation action performing; 6), state transition; 7), iteration and updating until termination condition is satisfied. By the initiative recommendation method, consumption dead areas brought by information overload is favorably reduced in the large scale data environment, and screening and rapid positioning of mass information can be realized.

Description

technical field [0001] The invention relates to the application of a multi-agent active recommendation method of a Q reinforcement learning algorithm to e-commerce of agricultural materials and commodities, is an application of intelligent decision-making in e-commerce, and belongs to the technical fields of artificial intelligence and automatic control. Background technique [0002] Agricultural production materials mainly include chemical fertilizers, pesticides, agricultural film, seeds, feed, agricultural machinery and other major commodities, which are important agricultural production factors. Among them, agricultural materials such as seeds, chemical fertilizers, and pesticides are key agricultural production materials, which are directly related to the production of agricultural products. Yield and quality are of special significance to ensuring agricultural production, increasing farmers' income, and maintaining national food security. [0003] Due to the special ch...

AI Technical Summary

Problems solved by technology

Obviously, these recommendation methods are essentially driven by the consumption behavior of customers, and cannot be applied to the active recommendation system in the field of agricultural material e-commerce driven by production demand.

At the same time, on large-scale datasets, in order to maintain the irregular migration of users and interests, this type of recommendation algorithm will bring huge storage and computing overhead, which will affect the response speed of the system to users to a certain extent.

Images